Implications of Risk Stratification for Diabetes Prevention: The Case of Hemoglobin A1c

Although glycated hemoglobin (HbA1c) has been widely recommended for the diagnosis of diabetes, considerable ambiguity remains about how HbA1c should be used to identify people with prediabetes or other high-risk states for preventive interventions. The current paper provides a synthesis of the epidemiologic basis and the health and economic implications of using various HbA1c-based risk-stratification approaches for diabetes prevention. HbA1c predicts diabetes and related outcomes across a wide range of HbA1c values. However, the authors estimate that, among U.S. adults, the top 15% of the nondiabetic HBA1c distribution (HbA1c of 5.7%–6.4%) accounts for 47% of diabetes cases over 5 years, and the top 30% (5.5%–6.4%) accounts for about 70% of cases. Although this clustering of eventual cases at the high end of the HbA1c risk distribution means that intervention resources will be more efficient when applied to the upper end of the distribution, no obvious threshold exists to prioritize people for preventive interventions. Thus, the choice of optimal thresholds is a tradeoff, wherein selecting a lower HbA1c cut-point will lead to a higher rate of eligibility and health benefits for more people, and a higher HbA1c cut-point will lead to fewer cases of diabetes prevented but greater “economic efficiency” in terms of diabetes cases prevented per intervention participant. Selection of optimal HbA1c thresholds also may change with the evolving science, as better evidence on the biologic effectiveness of lower-intensity interventions and effects of lifestyle interventions on additional outcomes could pave the way for a more comprehensive, tiered approach to risk stratification.